Analysis, pipelines, machine learning and science.
Turn code, docs and media into a queryable knowledge graph with communities and paths.
Reads public posts, comments, transcripts and market data from major platforms, read-only.
Estimates metabolic fluxes from carbon-13 isotope tracing data with fitting and diagnostics.
Forecasts univariate time series zero-shot with Google's TimesFM, with optional covariates.
Builds persistent agent memory and knowledge graphs with graph-based retrieval.
Reads A-share stock, financial, and valuation data through local runtime interfaces.
Inspects DBX database connections and runs bounded read-only queries and schema docs.
Draws bar, line, scatter, pie, and time series charts from data.
Checks A/B results for significance and sample size, then recommends a next step.
Runs the per-room calibration pipeline that trains small RuView sensing specialists.
Plans ablation studies once main results support a claim, for paper submission.
Loads SaaS data into DuckDB with dlt and builds a Wren semantic project for SQL.
Searches Zhihu through its API for answers and articles on a topic.
Discovers connected data sources, adds connectors, and probes tables read-only.
Search and summarize a local Discord archive, including DMs and sync status.
Analyzes cause-and-effect chains, interventions, and counterfactuals in a knowledge graph.
Generates publication-quality architecture diagrams and data charts for ML papers.
Designs protein assay experiments and submits sequences to the Adaptyv Bio Foundry API.
Sets conventions and review steps for analytics SQL, reports and metric aggregation.
Queries the ShadowBroker OSINT platform for geospatial intel, monitoring and reports.
Fetches A-share quotes, filings, financials, news and macro data with code.
Searches and reads research papers and their code through alphaXiv tools.
Compares inflation, rate and bond yield histories across countries, keeping sources.
Makes publication-ready matplotlib figures with house style and vector export.
Fetches free A-share, US, HK, futures, macro and forex data from AKShare.
Submits and monitors batch inference jobs and their files from the command line.
Queries a WeKnora knowledge base with chat, ask or chunk search via the CLI.
Trims, standardizes and deduplicates messy spreadsheet data before analysis.
Drives cognee memory, dataset, config and migration commands from the terminal.
Queries stock, fund, futures and macro financial data in natural language.
Search Granola meeting notes and transcripts in a local archive.
Inspects Honcho workspaces, peer memory and session context from the CLI.
Looks up symbols, K-line data and workspace conversations via the alice CLI.
Pulls observations into a seat cache hourly and flags conflicting rules.
Handles time series classification, forecasting, anomaly detection and similarity search.
Search Notion pages and databases in a local archive and export Markdown.
Compares two categorical datasets by count, ratio and difference, with charts.
Pulls and checks evidence from images, documents and video without redundant inspection.
Annotates VCF variants and normalizes HGVS with public annotators and gnomAD frequencies.
Builds a local, data-backed dashboard with charts, tables and a sources panel.
Handles astronomy data: coordinates, units, FITS files, cosmology and WCS.
Predicts monomer and multimer protein structures with ColabFold and scores them.
Picks a suitable chart type from 26 options and generates a chart image from data.
Gives methods for training, scaling and recovering large language and vision models.
Finds and labels every LLM workflow in a repo so spend can be grouped by workflow.
Summarizes one paper quickly from AlphaXiv output, given a link or arXiv ID.
Builds and validates one board of linked Flint charts for a dashboard or comparison.
Adds units, enum meanings and metric definitions that a database schema cannot hold.
